Human Soluble TRAIL Secreted by Modified Lactococcus lactis Bacteria Promotes Tumor Growth in the Orthotopic Mouse Model of Colorectal Cancer

Human colon carcinoma cell line HCT116 Red-FLuc (Bioware® Brite Cell Line BW124318) was obtained commercially (Perkin Elmer, Waltham, MA, USA) and maintained according to the distributor’s instructions, in a 37°C humified atmosphere with 5% CO₂. Briefly, the cells were cultured in McCoy’s 5A (Gibco, Paisley, UK) supplemented with 10% fetal bovine serum (Gibco, UK) puromycin (2 μg/mL; Cayman Chemical, Ann Arbor, MI, USA), passaged twice a week with 0.05% trypsin (PAN-Biotech GmbH, Aidenbach, Germany) in EDTA (Eurx, Gdansk, Poland) and regularly tested for Mycoplasma sp. contamination by PCR-ELISA test (Roche, Mannheim, Germany).

Lactococccus lactis NZ9000 host strain was obtained from MoBiTec (Goettingen, Germany). Design and preparation of genetically modified L.lactis clones, harbouring secretion plasmid vector pNZ8124 (MoBiTec; designated as “L.lactis(“empty” vector)”) or their counterparts with the hsTRAIL-cDNA insert, containing the optimized codons to fit the codon usage pattern (codon bias) of L.lactis, placed downstream of the inducible promoter PnisA (“L.lactis(hsTRAIL+)”), were previously described (Ciaćma et al., 2018). Bacteria were cultured without aeration at 30°C, in M17 medium (BTL, Lodz, Poland), supplemented with 0.5% glucose (POCH, Gliwice, Poland) and 10 μg/mL chloramphenicol (Cm10; Sigma Aldrich, Saint Louis, MI, USA) to maintain the plasmid. Bacteria stocks were frozen in M17 medium, supplemented with 0.5% glucose, Cm10 and sterile 20% glycerol (Eurx) and stored at −80°C until further use.

Reinforcement of the principle of 3Rs was considered in the experimental design. For orthotopic model of human CRC, mice of the SCID (CB17/lcr-Prkdcscid/lcrlcoCrl) strain (Charles River Laboratories, Wilmington, MA, USA), homozygous for the severe combined immune deficiency spontaneous mutation Prkdc^scid with the lack of mature B and T lymphocytes, as described previously (Bosma et al., 1983; Bosma 1991), were chosen. The model was based on human HCT116 Red-FLuc cell line with luciferase activity due to their transfection with luciferase gene from Luciola Italica. Orthotopic implantation of human CRC cells was commercially ordered in Oncodesign (Dijon, France), according to the established protocol. 51 females at 6–8 weeks of age with orthotopically developed CRC were used for the experiment. Mice were maintained in authorized animal facility of Oncodesign and local facility at the Jagiellonian University Medical College (Krakow, Poland). Mice were housed in dedicated housing rooms under aseptic and controlled laboratory conditions: 12/12 h light/dark cycle, room temperature 20–22°C, humidity 45–55%, HEPA filtered air, group housing, access to food and water ad libitum. Animals were monitored daily for signs of distress and health status. All experiments were performed in accordance with the European Union Directive (2010/63/EU), Polish and French law and were approved by the local Ethical Committees (Poland: 202/2018; France: CNREEA approval no.91).

To monitor the growth of HCT116 Red-FLuc-tumors, the mice were subjected to bioluminescence imaging (BLI). Ten minutes prior to imaging, the mice received intraperitoneal (i.p.) injections (100 μL, 125 mg/kg) of XenoLight D-luciferin (Perkin Elmer) in phosphate-buffered saline (PBS; Corning, New York, NY, USA), were anesthetized with 1–3% isoflurane and placed in the dark chamber of Ami HT. Detection and quantification of the photons, emitted from the bioluminescent tumors, were performed using Ami HT (Spectral Instruments Imaging, Tucson, AZ, USA). To optimize the signal level to avoid saturation of the light detector, the exposure time was adjusted to the amount of luciferase activity and luciferin pharmacokinetics, finally reaching 5 s. Images were analysed by manually defined regions of interest, using Aura Imaging Software (Spectral Instruments Imaging). Bioluminescence data were shown as Photon flux, defined as average number of photons per second, that radiate from the mouse in a unit area (1 cm²) and unit angle (1 sr). When all tumors exceeded 1.14 × 10⁵ [photon/s/cm²/sr], the mice were randomized according to the similar mean value of the signal level in each experimental group, divided into six experimental groups (n=8) and received the first dose of treatments (Day 0). The number of mice/experimental group was defined according to previous studies and published evidence. Mice were imaged along abdominal view at the following days: 0, 1, 4, 8, 13, 17, 22, 28, 32, 38.

The broth cultures of L.lactis(hsTRAIL+) and corresponding negative control (L.lactis(“empty” vector)) were prepared, as previously described (Kaczmarek et al., 2021; Ciaćma et al., 2018). The bacteria were administered by gastric gavage, at a dose of 2 × 10⁹ colony forming unit (c.f.u.), in a volume of 100 μL of lactose-free skimmed milk (10%), supplemented with ZnSO₄ (100 μM; Linegal Chemicals, Warsaw, Poland), nisin (50 ng/mL; MoBiTec) and aprotinin (2 μg/mL; Bioshop, Burlington, Canada), used as vehicle to increase the survival of L.lactis strain in the gastrointestinal tract before reaching the tumor site (colon) (Kos et al., 2000). All treatments with the bacteria were performed once a day, at the same time, for 37 consecutive days. Induced cultures of L.lactis were freshly prepared each day of the therapy. Simultaneously, the control group of mice (“Mock control” or “Control” (CRC)) were treated with vehiculum only, administered by gastric gavage at the same volume as bacteria to maintain the same experimental conditions.

Stock solutions of MetF (Sigma Aldrich) were prepared in PBS, sterilized by filtration through a 0.22 μm filter (Carl Roth, Karlsruhe, Germany) and stored at 4°C. The therapeutic was administered by gastric gavage at a dose of 250 mg/kg in 10% skimmed milk, once a day, at the same time for 37 consecutive days.

When all animals developed extensive metastases and the decrease of their initial body weight reached 20% (Day 38), the experiment was terminated. After BLI measurements, animals were euthanized by over-dosage of isoflurane, followed by spinal cord dislocation. Then, colons with tumors, liver, spleen and lungs were isolated. Due to the limited activity of luciferase after the animal’s death, the resected organs were immediately subjected to BLI. The colons and primary tumors were subsequently fixed in 10% formalin solution ON, for further hematoxylin-eosin (H&E) staining and immunohistochemistry (IHC) analysis. For IHC analysis, the slides (3.5 μm) of paraffin-embedded colons with CRC were treated with sodium citrate at 97°C for 35 min followed by incubation with appropriate antibodies. To examine the local production of TRAIL, slides of colon with CRC, were incubated with rabbit monoclonal anti-human TRAIL antibody (1:100; Cell Signalling Technology, Danvers, MA, USA). To verify the potential interaction between M2-macrophages and CRC cells, slides of colon and primary tumor were also incubated with rabbit anti-mouse mannose receptor (CD206) polyclonal antibody (1:4000; Abcam, Cambridge, UK) and rabbit anti-human monocyte chemoattractant protein-1 (MCP-1/CCL2) polyclonal antibody (1:1000; Thermo Fisher Scientific). All dilutions were performed in Dako Antibody Diluent (Dako, Glostrup, Denmark). Visualisation was achieved using Dako REAL^TM ENVison^TM Detection System (Dako) according to the manufacturer’s instructions. All slides were analysed in a blinded manner using OLYMPUS IX70 microscope (Olympus, Tokyo, Japan) and the CellSens Dimension software (Olympus). Measurements and calculation of the colon crypt depth were performed using paint.net (v. 4.1.6, Rick Brewster, Pullman, WA, USA).

The viability of orally administered bacteria in the location of CRC was assessed after homogenization of excised cecum with corresponding fragment of colon, followed by plating of the homogenates in the presence of selection marker (Cm10) and PCR analysis of plasmids isolated from the grown colonies. Briefly, randomly selected cecum was cut into small pieces with a sterile scalpel and transferred to the Eppendorf tubes. Homogenization was done in M17 culture medium in the presence of metal beads, at frequency 20 Hz for 2 min using a Tissue Lyser II (Qiagen, Hilden, Germany). The obtained homogenates were plated in a volume of 100 μL on M17 agar (1.5%) Petri dishes, in the presence of 0.5% glucose and Cm10. After 48 h of incubation at 30°C, the grown colonies were randomly isolated to prepare ON broth cultures. Plasmid DNA was isolated using the Miniprep DNA Purification Kit (Eurx) according to the manufacturer’s instructions for isolation of plasmids from gram-positive bacteria. Concentration of the isolated plasmids was measured using a Quawell Q500 spectrophotometer (Quawell Technology, Inc., San Jose, CA, USA). To verify the presence of the hsTRAIL-cDNA insert, PCR using primers 5′-TGGTACTCGTGGTCGTAGCA-3′ sense and 5′-GAAGCTTCGTGGTCCATGTC-3′ antisense, was performed. The PCR reaction products were subsequently separated on a 1.5% agarose gel with 0.5 μg/mL EtBr. Perfect ™ 100 bp DNA Ladder (Eurx) was used as a size marker. The gels were photographed using a Gel Logic 1500 Imaging System (Kodak, Rochester, NY, USA). To create the appropriate positive and negative controls, plasmids isolated from fresh stocks (stored at −80°C) of L.lactis(hsTRAIL+) and L.lactis(“empty” vector), respectively, were also examined.

Statistical analyses were performed using GraphPad Prism software (v. 6.00) and Microsoft Excel (v. 2019). Comparisons between multiple groups were performed using analysis of variance (ANOVA test) with post-hoc multiple comparisons by Tukey’s method. Data were presented as mean ± SEM. The statistically significant results were showed as *p<0.05, **p<0.01, ***p<0.001.

All graphics were created with BioRender.com.

Encouraged by the results from our previous studies with subcutaneous model of CRC (Kaczmarek et al., 2021), showing a significant retardation of the tumor growth in NOD-SCID mice after intra-tumoral secretion of hsTRAIL by L.lactis(hsTRAIL+) bacteria, here we examined the effect of L.lactis-delivered hsTRAIL in the orthotopic model of human CRC. To this end we used the model developed by implantation of human HCT116 Red-FLuc cells to cecum of SCID mice, allowing for monitoring the tumor size by BLI. Figures 1a,b summarizes an experimental setup for this therapy model. Simultaneous analysis of the pattern of bioluminescence distribution and number of photons emitted from the tumor showed a stimulatory effect of the L.lactis(hsTRAIL)-treatment on tumor development (Figures 1c,d). Daily administration of the hsTRAIL-expressing bacteria for 37 days significantly stimulated the growth of CRC, when compared to mock control (p<0.001) or animals treated with L.lactis(“empty” vector) (p<0.001). In this context it was clear that the observed effect was related to biological activity of continuously delivered hsTRAIL, but not to genetically modified bacteria by themselves (Figure 1c). The effect of co-treatment with MetF was also different, comparing to subcutaneous model (Kaczmarek et al., 2021). In this case, MetF given in monotherapy did not affect the tumor growth (p>0.5 compared to mock control), neither impeded the pro-tumorigenic activity of hsTRAIL in combined therapy with L.lactis(hsTRAIL+) (p>0.5), although the dose of MetF was intentionally doubled, expecting an enhancement of its activity (comparing to subcutaneous model, where MetF was used at the concentration of 125 mg/kg) (Figures 1c,d).

Fig 1.

In general, L.lactis species are sensitive for the action of digestive enzymes (Klijn et al., 1995; Berlec et al., 2015). In order to ensure the best survival of orally administered L.lactis(hsTRAIL+) or corresponding negative control, the bacteria were administered in a solution of skimmed milk in a dose of 2 × 10⁹ c.f.u. (intentionally doubled when compared to the subcutaneous CRC model, [Kaczmarek et al., 2021]) and their survival in the gut was assessed post mortem, by counting bacterial colonies grown from the cecum homogenates (Figure 2a). The presence of MetF allowed for better survival of both L.lactis(hsTRAIL+) and L.lactis(“empty” vector), indicating that this effect was not related to the ability of L.lactis-strain to produce hsTRAIL (Figure 2b). To exclude the contamination of the homogenates with other bacterial strains from the colon, we performed a PCR analysis for hsTRAIL-cDNA. Results obtained for the randomly picked colonies confirmed, that they originated from orally-delivered L.lactis(hsTRAIL+) (Figure 2c).

Fig 2.

IHC analysis of the colon tissue has documented a local secretion of hsTRAIL after oral administration of L.lactis(hsTRAIL+) bacteria, and this was not affected by MetF (Figure 3). It was shown for the orthotopic model of lung cancer that TRAIL can promote the tumor growth through the induction of MCP-1 secretion by cancer cells and accumulation of M2-like cells in the tumor (Hartwig et al., 2017). Therefore, in our experimental setup we examined the presence of M2-macrophages (defined as CD206-positive cells) and MCP-1 expression in the primary CRC tumor. In this context we found that CD206⁺ macrophages infiltrated solely tumors in mice receiving the hsTRAIL-producing bacteria, although their presence in colon tissue was detected in all treatment groups (Figure 4). This hsTRAIL-driven tumor infiltration of M2-macrophages however, did not correlate with the production of MCP-1 by HCT116 Red-FLuc cells in the tumor tissue - an increased secretion of MCP-1 by human CRC cells in the primary tumor was observed in all groups of animals receiving L.lactis bacteria, regardless of their ability to secrete hsTRAIL (Figure 5).

Fig 3.

Fig 4.

Fig 5.

To fully address the effect of the treatment on CRC-progression, we analysed changes in the colon morphology after 37 consecutive days of hsTRAIL administration. This was performed by H&E staining and the subsequent measurements of the colon crypts’ depth (Figure 6a). Oral administration of L.lactis(hsTRAIL+) resulted in a significant reduction of the crypts’ depth, when compared to healthy mice (Control (healthy); p<0.05), or to mock control (Control (CRC); p<0.01) (Figure 6b). The regular treatment with L.lactis(“empty” vector) did not affect the depth of the crypts, pointing on the role of systematic delivery of hsTRAIL, but not on L.lactis bacteria by themselves. However, this “devastating” action of hsTRAIL was abolished by MetF, suggesting its protective role on the colon crypts in CRC mice (Figure 6b).

Fig 6.